Вы находитесь на странице: 1из 30

International Journal of Pressure Vessels and Piping 81 (2004) 89118

www.elsevier.com/locate/ijpvp

Milestones in pressure vessel technology


J. Spence*, D.H. Nash
Department of Mechanical Engineering, University of Strathclyde, 75 Montrose Street, Glasgow G1 1XJ, UK

Abstract
The progress of pressure vessel technology over the years has been influenced by many important events. This paper identifies a number of
milestones which have provided a stimulus to analysis methods, manufacturing, operational processes and new pressure equipment. The
formation of a milestone itself along with its subsequent development is often critically dependent on the work of many individuals. It is
postulated that such developements takes place in cycles, namely, an initial idea, followed sometimes by unexpected failures, which in turn
stimulate analysis or investigation, and when confidence is established, followed finally by the emergence of codes ad standards. Starting
from the industrial revolution, key milestones are traced through to the present day and beyond.
q 2003 Elsevier Ltd. All rights reserved.
Keywords: Pressure; Boilers; Power; Analysis; Energy; Codes and standards; Hydrogen; Nuclear; Oil and gas; Pipelines; Milestones

Preamble
This paper was first presented as the Roy Nichols
Memorial Lecture at the 10th International Conference on
Pressure Vessel Technology, Vienna, Austria, on 7th July
2003.
Roy Nichols (Fig. 1) was born in Yorkshire, England, and
studied at Loughborough and Sheffield Universities. He
joined the Reactor Materials Laboratory at the UKAEA in
1956 and became Deputy Head in 1971 and later Head of
Laboratory. He retired from that post in 1985 and sadly died
in 1999. He played a major role over many years in the safety
aspects of nuclear reactors in the UK. A major strength was
his knowledge of problems associated with structural
integrity in different countries, as was his conviction of the
need for international co-operation and an internationally
consistent approach. Another significant contribution was his
involvement with the International Journal of Pressure
Vessels and Piping where he acted as Editor until he died.
In the late 1960s, he invited one of the authors (JS) to serve on
the Pressure Systems Group (PSG) committee of the
Institution of Mechanical Engineers in London, where he
served, initially under Roys chairmanship, until he became
President in 1998. The other author (DN) presently serves as
Vice-Chairman of the Pressure Systems Group.
In passing we would like to take the opportunity to
mention another colleague, Professor Thomas Jaeger who
initiated the Structural Mechanics in Reactor Technology
* Corresponding author.
E-mail address: j.spence@mecheng.strath.ac.uk (J. Spence).
0308-0161/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ijpvp.2003.11.002

(SMiRT) series of conferences which started in Berlin in


1971. Roy Nichols was also involved in the organisation of
SMiRT. These two men, Roy Nichols and Tom Jaeger, were
men of vision. It is one of the great satisfactions in life to be
able to look back and give credit to the contribution of
people who have influenced ones thinking and career.
1. Introduction
There are few milestones which one can identify at a
specific point in time and space. Rather, ideas and concepts
are generated and then they mature or develop. Therefore,
milestones here should be understood in the sense of
important developments but that would not have provided
such an attractive title. Of course, the title is used simply as
a peg on which to hang a few ideas drawn from history and
some personal experiences.
There is a cycle that sometimes operates as follows:
Firstly, there is the initial idea and its development in
products or new theoretical work, which collectively we
may consider a milestone.
Sometimes, there are unexpected failures.
This stimulates investigation or analysis or more detailed
analysis leading to better understanding.
Finally, when confidence has been established, we have
the emergence of codes and standards.
Of course things do not always happen in that order and
sometimes economics plays a major role, but for the purpose

90

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 3. James Watt (1736 1819).


Fig. 1. Roy Nichols.

of this paper, we will broadly use that sequence for


convenience.
2. Milestone: the Industrial Revolution
By any measure, the industrial revolution was a major
milestone. Engineering in general fuelled the industrial
revolution from the mid 18th century but it was steam that

Fig. 2. Newcomen engine.

provided the necessary power sources and enabled the


engineering achievements. Thomas Savery patented the first
crude steam engine in 1679. Thomas Newcomen in
Dartmouth developed the Newcomen engine, a more
practical design, Fig. 2, in 1712 but it was still quite
inefficient. It was used mainly for pumping water from coal
and tin mines.
A model of the engine was given to James Watt, Fig. 3,
for repair by Professor John Anderson (the founder of the
University of Strathclyde). Watt (1736 1819) with his
separate condenser is credited with improving the steam
engine (Fig. 4) to make it the workhorse of the industrial
revolution [1]. A simple memorial boulder in Glasgow
Green commemorates his inspiration (1765). Later, the
Boulton Watt partnership produced thousands of engines

Fig. 4. Watt engine.

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

91

from their factory in Birmingham. Iron founding companies


were quick to turn their expertise to steam cylinders and
before long these and associated items dominated their
output. Although these engines were static, they were
portable and more versatile than waterpower.
It was left to others to utilise the idea for propulsion [2].
Richard Trevithick in 1801 invented a steam road carriage
and took it to London in 1804, but this form of transport
never became popular. He then developed the first steam
railway locomotive, a version of which was built at
Gateshead in the North-east of England around 1804
1805. This region, with its network of existing rail wagon
ways and plentiful supply of coal would become a proving
ground for the new technology. Local engineers and colliery
owners developed commercial railway locomotives. The
figureheads for this movement were George Stephenson
(1781 1848) and later his son Robert Stephenson (1803
1859). Their Rocket (Fig. 5) was the culmination of design
experience up to 1829, when trials on the Liverpool
to Manchester Railway determined their professional
ascendancy. George Stephenson became the first President
of the Institution of Mechanical Engineers in 1847. Within a
few decades, by the late 19th century, all of Britain had been
opened up to railway travel. Similarly, steam ship propulsion was demonstrated by John Fitch in USA (1787) and by
Symington in Dumfries, Scotland (1787/1788) although the
idea had been proposed as early as 1736 by Jonathan Hulls
who suggested and patented the idea of using a Newcomen
engine to pull ships in and out of harbour. The original

painting by Hulls hangs in the Presidents Office in the


Institution of Mechanical Engineers, London, although it is
not certain that the concept ever came to fruition. By 1791,
there was a service established between Philadelphia and
Burlington, New Jersey. Later (1801), the Charlotte
Dundas was trialled in the River Forth and Henry Bells
Comet (Fig. 6) went into service on the Clyde in 1812, the
first passenger steamship on European waters. It started a
tradition, which was popular into the mid 20th century.
Indeed the Waverley ocean going paddle steamer still
operates on the Clyde today. The first steamship on the
Thames in London was the PS Margery which was built by
Denny of Dumbarton with engines from Cooks in Glasgow.
Thus, industry was freed from the restriction of
waterpower and roared into the industrial revolution in a
cloud of steam generated by coal, powered by the steam
engine [1]. Behind all these, developments lay an unsung
hero, the humble boiler. Newcomen, Watt, Trevithick,
Stephenson, etc. developed pressure vessels and pressure
systems technology from practical experience. The early
boilers were essentially atmospheric but gradually pressures
increased. There were no safety inspectors but as responsible engineers, they attempted to ensure safety and their
reputation through experimentation and testing.
In passing, it is worth noting that the practical
experimental approach to engineering development allowed
British engineers to be extremely creative during the 18th
and 19th centuries. Equally, the lack of a scientific
foundation was possibly the reason for the erosion of that
pre-eminence in the second half of the 19th century when
other countries, notably in Europe, had a more coherent and
planned approach to the application of new technology.
However, Scottish universities were teaching mechanical
science in the mid-18th century and James Watt and his
contemporaries benefited. It was well into the 19th century
before engineering became accepted as an academic subject
in English universities.
In the meantime, great strides were being made in
mainland Europe [3]. The Bernoulli brothers (1654 1705),
Euler (1707 1783), Lagrange (1736 1813), Coulomb
(1736 1806), Navier (1785 1836) and many others
contributed to the development of infinitesimal calculus

Fig. 5. Stephensons Rocket.

Fig. 6. Henry Bells Comet.

92

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

and an understanding of mechanics of elastic bodies which


laid the theoretical basis for future developments.
2.1. Boiler failures and the initial development
of design rules
With such rapid development and little theoretical basis,
it was inevitable that there would be failures. Boiler failures
were particularly common. These are well-documented in
the UK and the USA although they were also common in
Europe [4]. In the UK in 1817, a Select Committee was set
up To consider means of preventing mischief of explosions
on steam boats to the danger or destruction of His Majestys
Subjects and the first Factory Act was passed in 1833. In
1854, the Manchester Steam Boiler Assurance Company
was set up with the purpose of both inspecting and insuring
boilers against damage arising from explosion.
In Germany, legislation in industrial safety began in
Prussia in 1831 with the Supreme enactments concerning
the plant and use of steam engines and the Prussian
General Industrial Code in 1845 which gave general
requirements. In 1856, the Law on the operation of steam
boilers prescribed regular external inspections and hydrostatic testing. The Factories Act of 1872 stipulated
requirements for the supervision and control of steam
boilers but these could be exempt from official tests if the
operator was a member of an association whose full-time
inspection engineers conducted regular tests and inspections
on the steam boilers. This laid the foundation in Germany
for the formation of steam boiler supervision, inspection
(and insurance) organisations [5]. Over a 25-year period at
the end of that century, the UK Board of Trade investigated
1871 boiler explosions which had led to 732 deaths but it
was not until 1934 that Lloyds Register of Shipping issued

the first design rules, Tentative requirements for fusion


welded pressure vessels. Experience was the major
contributor to the development of the early design rules.
To some extent, this is still the case although many of the
developments are based on analysis with validation testing.
Parallel developments took place in USA. As early as 1830,
the Franklin Institute of Philadelphia started to investigate boiler
explosions and made a number of recommendations. These
included recommendations that all boilers should have a proper
gauge, a glass water gauge, two safety valves and so on; they
also stressed the need for frequent internal inspections and low
level water alarms. The Steamboat Act of 1852 provided the first
simple rules and regulations for boilers. In 1889, the American
Boiler Manufacturers Association was formed but did not make
much headway due to conflicting commercial interests. An
important failure took place in Hartford on 2 March 1854, where
a boiler exploded in a factory killing 21 people and seriously
injuring 50 more. This explosion was a catalyst [6] for several
actions not least being the charter for a new hospital. In Hartford,
some businessmen formed the Polytechnic Club in 1857 with
the aim to avoid boiler explosions through good materials, fine
workmanship, careful operation and periodic inspection. It was
intimated that at that time, boiler explosions were occurring at
the rate of one every four days.
On 27 April 1865, there was another spectacular
explosion on the Mississippi steamship Sultana, which
proved to be the worst ever boiler accident with the loss of
1238 lives, albeit more were lost through drowning (the ship
sank in 20 min) than from the explosion (Fig. 7). Some
members of the Polytechnic Club put together a bill for the
incorporation of The Hartford Steam Boiler Inspection and
Insurance Company (HSBC) which came into being in
1866. Other companies have come and gone but this one was
to remain and it still resides in Hartford today (Fig. 8). At One

Fig. 7. The SS Sultana.

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

93

Technology was in its infancy yet the humble boiler lay at


the heart of the steam raising process. In this case, the cycle
of development, failure and simple rules took place with
virtually no analysis being performed.
3. Milestone: electricity

Fig. 8. The Hartford Logo.

State Street, which is the company headquarters, the 20th


floor houses the Polytechnic Club, which also still exists,
although its interests today are somewhat removed from
boiler explosions. It was a milestone of a kind for one of the
authors (JS), accompanied by a colleague Professor Alwyn
Tooth, to dine in the Club on the occasion of delivering a
course on Pressure Vessel Design at Hartford in 1993.
In due course, HSBC began to offer advice and issue its
own guidelines to boiler owners. The first rules in the USA
were adopted in Massachusetts in 1907 after two serious
explosions in shoe factories10 March 1905, 58 killed, 117
injured and 6 December 1906 in Lynn, where the factory was
completely destroyed. The Governor of Massachusetts set up
a committee, which produced basic rules although they were
only a few pages long. The President of ASME in 1911, Col.
E.D Meier was also the head of a boiler company. He was
instrumental in persuading the ASME Council to set up a
committee to formulate a standard specification for the
construction of steam boilers and other pressure vessels and
for the care of the same in service. The committee adopted
and modified the Massachusetts rules. Council approved them
on 12th March 1915 being 114 pages long with a detailed 28
page index [7]. This was the first edition of Section IPower
Boilers of the ASME Boiler and Pressure Vessel Code. In
1921, Section III Boilers of Locomotives was issued (later
incorporated into Section I) and in 1925, Section VIII
Unfired Pressure Vessels became available. At this point
of time, the American codes seemed well ahead of others.
Explosions still happen. On 25th May 2003, the cruise
ship SS Norway, while docked in Miami, suffered an
explosion in the boiler room following a steam leak. Four
crew were killed and 20 injured.
Recently, Professor Xue-Dong Chen of Hefei General
Machinery Research Institute, China, has drawn the authors
attention to the Chinese statistics for pressure related
failures [9]. In 1999, out of a total of 138 explosions and
150 accidents with 131 fatalities, 31 explosions and 123
accidents resulted from boilers with 21 fatalities. In 2001,
the situation was worst with a total of 284 fatalities; in
boilers, pressure vessels and pressure piping accidents, 158
people died and 342 were injured. Many of these were due
to illegally produced equipment and inadequate in-service
inspection. There are still lessons to be learned.
The industrial revolution was undoubtedly a
major milestone in all manner of ways. Pressure Vessel

The coming of electricity was surely a milestone in every


sense of the word. Although Michael Faraday is given credit
for generating electrical current on a practical scale, the
development of electricity was a truly international effort.
The names associated with some of the electrical units tell
the story, Volta, Faraday, Tesla, Watt (again), Ampere,
Ohm, etc. [8]. Edison and Swan made other significant
achievements especially with their incandescent filament
lamp. In 1881, Godalming, on the River Wey halfway
between London and Portsmouth, became the first town to
be lit by electric light (the bid for electricity was 15 less
than that for the continuation of gas lighting so economics
had a part to play). In September 1882, the first New York
Street was lit by electric lamps using direct current. In
1890s, George Westinghouse with Nicola Teslas help
persuaded American society that the future lay with
alternating current. Similar pioneering work was taking
place in Hungary where Otto Blathy, Miksa Deri, Karoly
Zipernowsky are credited with developing the economical
transmission and distribution of light from electricity over
long distances based on alternating current. On 12 January
1882, Thomas Edison opened the Edison Electric Light
Station at Holborn Viaduct in London. Six weeks later, the
Hammond Electric Light Company opened Brighton power
station, which claimed to be the first permanent and viable
public power supply. The early applications combined
Edisons generator with Watts reciprocating steam engine;
however, the development of the steam turbine based on the
Rankine cycle (William J M Rankine of Glasgow University
in 1859) soon led to the turbine alternators used in the
emerging power stations. In Britain, the National Grid
System was started in 1928 and completed 17 years later.
In fact, it was only on the 3rd of June 2003, that the last
village in the UK, Cwn Brefi in Wales (with 12 households)
was connected to the National Grid.
3.1. Boilers and their development
Once again, behind these significant developments lay
the humble boiler. Not so humble now since it too was
developing. In 1867, George Babcock and Stephen Wilcox
patented the Babcock and Wilcox Non-Explosive Boiler.
Babcock was a founder member of ASME in 1880. By
1890s, the Babcock boiler (Fig. 9) had become essentially a
standard product. It had forged steel drumheads, crossboxes, nozzles and tube headers and had gained a reputation
for reliability. The first electric lighting systems were
powered by Babcock boilers. Edison said, The Babcock
boiler is the best boiler God has permitted man yet to make.

94

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 9. The Babcock boiler.

The fundamentals of the designan overhead steam drum,


a mud drum at the lowest point and the furnace transferring
heat to a cluster of inclined water tubespersisted in the
design of steam plant for the next hundred years.
In the context of the ICPVT-10 venue (Vienna), it was
interesting to discover that an early Babcock achievement
was the successful development of a forging technique for
making steel headers [10]. The first European order
incorporating this innovation was for the Vienna opera
house (Fig. 10), where six Babcock boilers supplied steam to
power electric lighting. The price delivered to Vienna was

530 (800 Euro) per boiler. The opera house was originally
opened in 1869 and the boilers installed in the late 1880s.
Fig. 11 shows the Holborn Viaduct station mentioned earlier.
A Babcock boiler can be seen in the sub-basement; it supplied
steam to a reciprocating steam engine on the floor above.
The boiler industry grew at a tremendous pace in the
second half of the 19th century driven largely by the
demand for steam. However, it was the coming of electricity
and in particular, the idea of distributing electric power from
a central source, that placed new requirements on the steam
boiler and stimulated its further development in the early
years of the 20th century. Small units gave way to larger
units with cross-type boilers, water-cooled furnaces,
economisers and superheaters. Coal remained the main
fuel for some time but by the 1930s, pulverised fuel firing
was established. Pulverising mills became standard features.
Many times, one of the authors (JS) and colleagues worked
inside these mills in cramped, hot conditions to attach strain
gauges to the loading arms in an attempt to understand the
loading patterns and avoid fatigue failures.
A landmark in UK terms was passed in 1920 when a boiler
operated at a pressure of over 1000 psi (6.9 MPa). Traditional
riveted manufacture put a practical limit on the thickness of
the steam drum wall. Forged drums were very expensive and
limited in other ways. In 1932, Babcock developed a reliable
fusion welding process for ferritic materials and produced

Fig. 10. Vienna Opera House.

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

95

Fig. 11. The Holborn Viaduct Station.

the first fusion-welded steam drums to satisfy the Admiralty


and Insurance authorities specifications. This was probably
the single most valuable development at that time. It allowed
much higher pressures. Shortly afterwards, the cyclone steam
separator was developed which was also important. At high
pressure, it becomes increasingly difficult to separate steam
and water so the cyclone was invaluable in assisting the
delivery of dry steam to the turbine. After the Second World
War, power station capacities steadily increased (Fig. 12),
with 60-megawatt (MW) units in 1946, 120 MW units in
1952, 200 MW in 1956. Multiple units were common of
course often being six or eight in number. By the 1990s,
1200 MW was common. In the UK, the unit size is currently
restricted to 1300 MW in order to limit the damage to the grid
in the event of a sudden trip.
The growth of established companies and the expertise of
the engineering inspection and insurance companies helped
to dramatically reduce the incidence of boiler and
pressurised system failures although they were not
altogether eliminated. Likewise the experience accumulated
backed by relatively simple methods of analysis led to the
development of codes and standards in several countries.
A considerable amount of common material exists
between the main pressure equipment standards and indeed
there are also many parallels in their development and
history. The above has concentrated on boilers for
simplicity and because of the direct impact from the coming
of electricity. However, the same stimulation took place in

terms of other pressure equipment and indeed much broader


developments.
3.2. Further development of standards
As an example of the effort in standards, it is sufficient to
focus mainly on the British scene by way of example.
Tentative efforts to produce a British Standard specification
for fusion-welded vessels were made as early as 1939. A
serious start in the preparation of what became BS 1500 was
made in 1940. This resulted from approaches to BSI by the
Institution of Chemical Engineers. It is interesting that
economics rather than safety drove this approach. The
various BSI specifications of the day for welded pressure
vessels were rather more conservative than those issued
jointly by the American Petroleum Institute and the
American Society of Mechanical Engineers. This was

Fig. 12. Growth of station outputs.

96

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

having an unfavourable effect on the export trade. The first


draft British Standard ran to 94 pages and contained 56
graphs and sketches. Design stress tables were prepared but
were omitted from the first version as it was considered
politic to concentrate on the scope and format of the
proposed code (committees still disagree on this aspect
today). It became apparent that a series of supporting
specifications for materials would be necessary. This was
against BSI practice at the time. However, draft documents
were prepared for plates, forgings, tubes, etc.
On a historical point, in 1943, a member of the IMechE
Pressure Systems Group met in London with other
colleagues to discuss pressure vessel standardisation. The
meeting was organised by BSI. Over a period of a few
meetings, the format and requirements of a master pressure
vessel standard were set down. The notes of these meetings
show that later, BS 5500 [11] and the European UFPV
Standard [12] follow the lines proposed by this group for a
master code.
In November 1949, the BS 1500 Standard was published
as a provisional standard. It was later reviewed and issued
as BS 1500:1958 superseding the provisional version. Later
on, in 1965, BS 1515 [13] was published which included a
more detailed and advanced approach based on higher
allowable stresses. In the late 1960s, the Morrison Enquiry
into the Pressure Vessel Industry proposed that all the
pressure vessel interests of BSI should be under one
committee (PVE/-) to rationalise the activities. Roy Nichols
played an important role in this enquiry and the final report
is sometimes referred to as the Nichols Report. This led on
to the formulation of BS 5500, the master Pressure Vessel
Standard, which was first issued in 1976. This standard was
used widely in many industries and by many countries and
was updated annually and reissued every 3 years.
In BS 5500, there were two approaches to design. Firstly
Design-by-Rule, or Design-by-Formula, which involves
the use of formulae and rules to determine the vessel
thickness and keep the nominal stresses below the tabulated
values. Secondly, following the ASME Standard of that
time, Design-by-Analysis (DbA) was introduced; this
requires the use of analysis, usually stress analysis, to keep
certain calculated stresses below the allowable tabulated
values. Most national codes and standards have incorporated this twin approach to design to some extent. BS 5500
ceased to be a national standard with the advent of the new
European Standard for Unfired Pressure Vessels EN 13445
but still exists in the form of a design guide as PD 5500. This
aspect will be elaborated on later. Parallel developments
resulted in related standards for boilers and other pressure
equipment.
Thus, the coming of electricity and the emergence of
large centralised power stations stimulated the development
and growth of standards in the pressure systems area
resulting in many of the standards with which we are
familiar today.

4. Milestone: nuclear power


Nuclear fission was discovered in the 1930s and seemed
to hold much promise as an energy source. By the 1950s,
research in nuclear power was quite advanced in USA, the
UK, France, Canada and the former Soviet Union. In 1956,
Calder Hall, in Cumbria, England, the first nuclear power
station in the world to produce electricity commercially was
commissioned. Chapelcross, a similar station and the first in
Scotland followed in 1959. Thereafter in the UK, nine
additional Magnox stations were built. Many of these are
still in operation. France initially also concentrated on gas
cooled reactor technology. Between 1958 and 1966, they
built six nuclear power stations (natural uranium, graphite,
gas cooled, NUGG). Later from 1970, France concentrated
on Pressurised Water Reactors, two being authorised in
1970 and many more subsequent to the oil crisis in 1973.
The USA, of course, had concentrated on developing watercooled reactors, both boiling water reactors and pressurised
water reactors (PWRs). Russia also pursued an extensive
nuclear programme.
Fast breeder technology was not neglected but has not yet
been exploited. In the late 1950s, a small experimental
(15 MW) and later a prototype fast reactor (250 MW) was
built at Dounreay in Scotland. The pressure vessels, steam
generators, pipework and other equipment for the DFR were
supplied by Babcock and Motherwell Bridge supplied the
containment sphere (Fig. 13). At that time, the sphere was the
largest constructed in Europe, 135 ft diameter and 1500 tons,
although larger has been built in USA. Probably the greatest
challenge was the remoteness of the site and the adverse
weather conditions. Dounreay is now being decommissioned
but the famous dome will remain. Historic Scotland has
agreed to make it a listed building. Parallel developments
took place elsewhere at various times. A summary of fast
reactors built including some aimed at electrical output is
indicated in Table 1. Few of these were intended to produce
electricity commercially and only one or two are currently
operating. Nevertheless considerable experience has been
accumulated. USA also built specialised fast reactor-type

Fig. 13. Dounreay power station.

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

97

Table 1
Fast reactors
Demonstration

Prototype/commercial

France

Rapsodie

10 MW

(1967-)

Phenix
Superphenix

250 MW
1200 MW

(1973-)
(1986-)

Germany

KNK II

20 MW

(1978-)

SRN 300 (did


not operate)

280 MW

(1991-)

Japan

Joyo

(1978-)

Monjo

300 MW

(1994-)

Russia

BR1
BR5
BOR60

BN 350
BN 600

150 MW
600 MW

(1972-)
(1980-)

12 MW

(1955-)
(1959-)
(1969-)

UK

DFR

15 MW

(1959-)

PFR

250 MW

(1974-)

USA

EBR-1
Enrico
Fermi
EBR-2

0.2 MW
60 MW

(1951-)
(1963-)

20 MW

(1963-)

facilities. One of the authors (JS) was privileged to visit Joyo


in 1977 just before it went critical.
The senior author joined Babcock and Wilcox Research
Station in 1959. It was the hey day of nuclear power in the
UK and over the next few years, he was involved with four
of the main stations being built, namely Hinkley Point (A),
Trawsfynnyd, Sizewell (A) and Wylfa. Almost immediately
he was involved in the pressure acceptance test of the first
Hinkley Point reactor sphere, 67 ft (20.42 m) diameter, 3 in.
(76.2 mm) thick (Fig. 14). This required over 300 strain
gauges and involved a pneumatic test which lasted several
days [14]. Typical strain gauge layouts at the internal
support skirt and the charge tube nozzles are shown in
Figs. 15 and 16. The site was cleared of personnel during

the test and the testing team sheltered behind the concrete
shield of the other reactor. Eighteen months later, the
section head had departed for academia and the senior
author (JS) found himself in sole charge of testing the
second reactor vessel at the ripe old age of 24.
A few years later, during the test of the reactor vessels at
Trawsfynydd in North Wales [15], the team was becoming a
little blase. Then, something went wrong. Some of the strain
gauges in the skirt region started to behave somewhat
strangely as if the skirt region of the vessel and support were
yielding and deforming at rather low stress levels (about
30% of what was expected). Fig. 17 shows the strain
behaviour schematically. The team had also been involved
in the site stress relief of these reactor vessels and there was

Fig. 14. Hinkley Point power station.

98

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 17. Unusual strain behaviour at Trawsfynydd.


Fig. 15. Strain gauges at shell/skirt junction.

a worry that stress relief had not been carried out properly.
The site management team (CEGB, Government officials,
inspectors, consortium representatives) were in constant
session and were having visions (or nightmares) of the
sphere slowly collapsing over the skirt.
Unfortunately, the gauges concerned were inside the
vessel and could not be examined without stopping the test
and reducing the pressure to zero. The advice given was to
stop in order to examine the gauges. The management team
was composed of senior people but they were not expert in
strain gauge technology and they tried to get agreement to
continue. After a few more pressure increments and further
discussions, they conceded and depressurised the vessel.
Once inside, it was discovered that the gauges
were wet; they happened to be exactly opposite the main
CO2 gas duct which was being used to pressurise the vessel.
When the compressors had started up, several gallons of
water lying in the coolant duct had been thrown over the
gauges. There were dehumidifiers on the system but no one
had thought to check for water lying in the ducts. Once the
area was dried out and the gauges replaced, the test
proceeded successfully. A full day was lost and all site
personnel had a days holiday on full pay plus bonus.
We were forcibly reminded of this incident recently
while watching a BBC2 TV dramatised documentary on the
NASA Challenger space shuttle that failed on January 28

Fig. 16. Strain gauges at a charge nozzle.

1986. The engineers did not want to launch because of the


low ambient temperature and anticipated problems with the
O ring seals (uncharacteristically, the temperature forecast
was 29 8F at Cape Canaveral). The managers used all the
usual managerial and political arguments to persuade the
engineer in charge. Finally they said It is time to take off
your engineers hat and put on your management hat; this
needs a management decision. The launch proceeded with
the disastrous consequences that we all know.
The lesson learned at Trawsfynydd, although it might not
have been articulated in this way at the time, was simply
there is never a time to take off your engineers hat.
Not everything was right, first time and there were
various failures that required modifications or redesign. In
1963, the Sizewell heat exchanger failure was as spectacular
as it was unexpected. Sizewell was similar to Hinkley in
concept except that the mitre bends and bellows units had
been replaced by piping which incorporated flexible smooth
bends, Fig. 18. In this case, the testing of the reactor sphere
was conducted by an insurance agency but the piping and
the heat exchangers were a Babcock responsibility. One of
the authors (JS) was involved in the testing of both.
These vessels were 22 ft 6 in. (6.8 m) diameter 94 ft
(28.6 m) long 21/4 in. (57 mm) thick in Mn Cr Mo V steel.
In normal operation, they resided vertically and were
pressurised with gas but they were tested hydraulically in a
horizontal position. The weight of the vessel was about
335 tons but increased to 1200 tons when full of water. One
of the vessels failed during the hydro test (Fig. 19). The senior
author was sent to site immediately and later was given the
job of fully strain-gauging the next one before testing. He was
a little worried that his superiors imagined that the strain
gauges would help hold the vessel together! The official
report on the failure was published [16] but the situation was
complex. During the hydrostatic tests, the boilers were
supported on temporary supports. These consisted of eight
pairs of steel chocks (Fig. 20). Four boilers had been tested
satisfactorily in this fashion. The next test had proceeded to a
pressure of 435 440 psig before leakage at a pump gasket.
The test pressure was 450 psig (3.1 MPa). A few days later at
virtually the same pressure, the failure took place. Failure
origins were identified at thermal sleeve fillet welds but no
significant pre-existing defects were found. Inferior impact
properties in the failed plates were attributed to the passage of

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

99

Fig. 18. Sizewell power station failures.

the brittle crack and this was substantiated by test work. This
material exhibited a Charpy V impact transition temperature
close to ambient in some conditions. However visual
evidence led to the conclusion that the main cause of failure
was that one of the chocks had failed causing a shock loading
on the vessel and additional dynamic stress. A more modern
analysis and explanation were given much later [17].

It was not just in nuclear plant of course that failures


surprised us. There were several well-documented failures
from this era [18]. The Cockenzie boiler drum was one of
these. This is mentioned partly because Roy Nichols was
a member of the Board of Inquiry and the Technical Sub
Committee. The official report [19] records that on 6 May
1966, while approaching the full hydraulic test pressure,
the drum failed at a pressure of 3980 psig (27.44 MPa)
with two longitudinal fractures 16 ft (5 m) long (Fig. 21).

Fig. 19. Sizewell heat exchanger failure.

Fig. 20. Support system.

100

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

4.1. Analysis capability

Fig. 21. Cockenzie boiler drum failure.

In this case, a pre-existing crack was discovered. The


crack was at an economiser nozzle and was 13 in.
(330 mm) long and had penetrated to a maximum depth
of 3.5 in. (89 mm), Fig. 22. It was established that it had
occurred during manufacture and was associated with
stress cracking in highly strained situations during stress
relief treatment.
Interestingly the boiler had already been to higher
pressures. First to 4100 psig (28.3 MPa) and then unofficially to the full test pressure of 4163 psig (28.7 MPa).
Undesirable as such a crack might be it was concluded that
the presence of an angle bracket welded to the shell
immediately adjacent to the nozzle fillet probably contributed to the failure. The bracket can be seen in Fig. 22.
The importance of inspection after heat treatment was
highlighted.

The rapid development of the nuclear industry and the


various failures provided an enormous stimulus to analysis
capability. This took place initially in USA but rapidly
spread to Europe, Japan and elsewhere. The first International Conference on Pressure Vessel Technology, held in
Delft in 1969, was timely; it allowed the analyses successes
of the day to be demonstrated. Shell theory (more later),
which had been an academic toy, rapidly became an
essential analysis tool. The DbA approach, which was
conceived in America [20] during this period, was very
much based on the assumption that shell-type analysis
would become the norm. Indeed the stress categories within
DbA pose considerable problems with almost any other type
of stress analysis. Of course, DbA included other
approaches such as shakedown and limit analysis but it is
only relatively recently that these have received attention as
a serious route to satisfy DbA requirements. Although
trailed in advance, the Design by Analysis approach became
generally available through ASME Section III Nuclear
Vessels issued in 1963, later to become Nuclear Power
Plant Components in the 1971 Edition [21]. It is also
available in ASME VIII Div 2. More will be said about
analysis capability later.
Nuclear power is not popular currently. Much of the
existing plant is due to be run down and not replaced. In the
UK, the situation is particularly uncertain. Nevertheless,
despite various difficulties and some notable accidents,
failures (rather less than steam boilers!) and public
disenchantment, it has been a real milestone in technical
development. In this context, Three Mile Island and
Chernobyl are considered accidents rather than technical
failures. Approximately 18% of the electricity supply
worldwide, 358,000 MW (2001 data), comes from nuclear,
roughly the same as hydropower. In several countries, the
percentage is much higher, see Table 2.
4.2. Case study: IPIRG pipe system test

Fig. 22. Cockenzie: pre-existing crack at nozzle.

The International Piping Integrity Research Group


involved 10 countries with a budget of US$11 M for the
first programme. The objective of the programme was
to provide a validated basis for the assessment of
leak-before-break under dynamic loading, specifically
for primary piping in nuclear plant. The programme
included many tests including piping system tests. The
programme was managed by the US Nuclear Regulatory
Commission and undertaken by the Battelle Institute,
Columbus, Ohio, USA. The test layout for the piping
tests is shown in Fig. 23. A part through defect was
subjected to a representative combination of inertial and
displacement controlled loading with a ramped sinusoidal
waveform at about 3 Hz. The tests were at PWR primary
circuit pressure and temperature conditions. The test
section had baffles with a small hole each end of the test

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

A number of interesting aspects were noted, including:

Table 2
Nuclear reactors world-wide c2001
European reactors
and TWh
Armenia
Belgium
Bulgaria
Czech Republic
Finland
France
Germany
Hungary
Lithuania
Netherlands
Romania
Russia
Slovakia
Slovenia
Spain
Sweden
Switzerland
Ukraine
United Kingdom
Total

Other reactors
and TWh
1
7
6
4
4
58
20
4
2
1
1
29
6
1
9
11
5
16
35
221

2.08
46.6
14.53
13.36
22.07
375
160.4
14.1
9.86
3.4
4.81
110.9
13.12
4.49
56.47
70.1
70.1
67.35
91.19
1150

101

Argentina
Brazil
Canada
China
India
Japan
Mexico
Pakistan
Republic of Korea
South Africa
United States
Total

2
1
14
9
11
53
2
1
16
2
104
215

6.59
3.98
70.4
14.1
11.45
306.9
9.56
0.69
97.82
13.47
719.4
1254

World total reactors


World total TWh*

436

2404

TWh Terawatt hours 1000 billion Watts hours, or billion kWh.


Source: International Atomic Energy Agency.

section so that when the leak occurred, the pressure


dropped rapidly at the test section.
The final test in the series was on austenitic stainless
piping provided by France. In order to better demonstrate
the effect of pressure on the leaking crack, it was considered
that the baffles should be removed. The test ran normally
and all the required data was generated. However, when the
crack started to leak, the pressure at the location did not drop
rapidly as there were no baffles. Consequently the pressure
loading on the pipe system was maintained leading to a
catastrophic failure. Part of the building was damaged,
missiles were produced and the test facility severely
damaged.

Fig. 23. IPIRG piping system layout.

The development of a plastic hinge development and


subsequent failure.
The test section was projected outside the building and
impacted on a number of pipes stacked outside. These
pipes themselves were projected a distance of between
60 and 150 m.
The steel cables across the test section to prevent separation
failed (it transpired that the cables replaced after a previous
test were 12 mm instead of 25 mm diameter).
The release of pressure inside the building blew out
upper section wall panels.
In summary, it made rather a mess!
The test produced a vast catalogue of information far
more than that yielded from the originally designed
experiment. It is unlikely that the test would have been
undertaken as a designed experiment if the consequences
had been anticipated. A video clip courtesy of John
Darlaston is available illustrating the tests online at www.
psg.strath.ac.uk/videos.

5. Milestone: oil crises


By the late 1960s, many economies around the world had
become dependant on countries exporting fossil fuels. The
oil crises of 1972/1974 forced a rapid reconsideration of
sources of energy. Fortunately for the UK and other countries
bordering on the North Sea, oil and gas reserves were
discovered just about that time. In 1973, the Arab oil
embargo forced the price of oil from $2.90 to $11.65 per
barrel; oil had entered the realm of power politics. The
Ecofisk gas field was discovered in December 1969, as it
happens just after the first ICPVT. The Forties field was
discovered in 1970 and proved to be the largest find in the UK
sector of the North Sea. Production started in 1974 and to
date, it has produced more than 2 billion barrels, equivalent to
15% of all UK production. Forties was sited in ca. 130 m
water depth, which was considerably deeper than experienced elsewhere in the world. At that time, most offshore
production in the Middle East or the Gulf of Mexico had been
in relatively shallow water. It was decided to bring the oil
from Forties field ashore by a 170 km long, 32 in. (813 mm)
diameter pipeline to land at Cruden Bay near Aberdeen.
British Petroleums previous experience in the West Sole
Field was of 90 ft water depth (27 m) (Fig. 24). In deeper
water, there were a number of new problems. A schematic
arrangement for pipe laying at that time is shown in Fig. 24.
The pipe is subject to bending at the overbend as it leaves
the stinger and then to reversed bending coupled with
external pressure due to the water depth at the seabed. Pipe
buckling may occur due to the combinations of loading
(including the inherent dynamic loads involved in the
process). If the pipe buckles, there is a second potentially

102

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 26. Hyperbaric chamber for pipe tests.


Fig. 24. Pipe laying schematic.

more important problem. In deeper water, a buckle, once


initiated, may propagate due to the external pressure alone;
if that happens, it propagates at approximately the speed of
sound in the metal. Forties Field was in a depth of water
where buckle propagation was possible and it was necessary
to insert buckle arrestors in the line. An illustration of an
actual pipe being laid in the North Sea is given in Fig. 25. A
pipe buckle in the North Sea cost approximately 1 million
to repair at that time.
Work was conducted [22] for BP on model pipes, ca.
150 mm diameter, but of the correct grade pipeline material
under combined bending and external pressure, up to
pressures well in excess of those encountered on Forties
Field. The idea was to map out combinations of conditions
that might cause buckling and also to investigate the
efficiency of different types of buckle arrestor [23]. The bulk
of the results remain confidential. The arrangement, which
essentially comprises a four-point bending-rig inside a
pressure vessel, is shown in Fig. 26. The bending load was
supplied by a pair of hydraulic rams mounted externally
with long push rods operating through glanded nozzles.
Incidentally, the (almost) full-bore closure on this 1 m
diameter vessel was a 32-in. pipeline flange identical to
those used on the Forties pipeline. A typical buckled pipe is

shown in Fig. 27. Simple arrestors, i.e. larger pipes slipped


on externally have, of course, their own buckling conditions
(Fig. 28) and if a heavy ring of pipe is used, the buckle can
flip into a different shape and traverse inside the arrestor
(Fig. 29) unless the arrestor is welded to the pipe. Of course,
buckles did occur including running buckles. However,
much analysis was conducted both in USA and Europe and
the phenomenon is now well understood. Pipelines have
subsequently been laid in much deeper water. BP sold the
Forties field in 2003 to the US company Apache. A video
clip courtesy of BP is available online that illustrates
pipelaying at www.psg.strath.ac.uk/videos.

Fig. 25. Actual pipe laying in the North Sea.

Fig. 28. Loose arrestors.

Fig. 27. Buckled pipe.

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 29. Flip flop buckle (stiff arrestor).

The oil crises and the development of the North Sea oil
and gas fields certainly stimulated analysis and research in
sub sea pipelines and resulted in greatly enhanced understanding of the buckling behaviour under laying and other
conditions. Specialist companies also developed. Usually
this work did not result in design codes since the companies
naturally guard their knowledge and experience carefully.
The oil and gas industries were and are major customers
for pressure equipment of various types both upstream and
downstream in refineries and chemical plant. These aspects
will not be discussed here.
6. Milestone: modern analyses
Modern analysis is so wide and varied that it is difficult to
capture it as a milestone-type event, yet in many senses, it
does encapsulate a milestone in pressure vessel technology.
One might ask the questions related to modern analyses
capability, what can it do? and how did we get here?. The
answer to the first question seems to be almost anything;
the answer to the second reveals an interesting step-by-step
(or mile-by-mile to keep to our topic) process.
It is salutary to remember that 300 years or ago at the end
of the 17th century, only four people were working in the
area of infinitesimal calculus [3]. These were Leibnitz,
Newton and the Bernoulli brothers. The first book on
Strength of Materials by Girard was about a century later.
Thereafter, a rich vein of talent across Europe made many
advances in the analysis and understanding of the mechanics
of elastic bodies. There was little in terms of pressure
systems although Fairbairn (1789 1874) conducted careful
experiments on tubes under external pressure (for boilers),
Grasshof developed a critical buckling pressure for tubes
and Winkler (1835 1885) identified the interesting behaviour of the bending of a thin pressurised tube. Rossmanith
[25] has reported that much earlier ca. 1650, Mariotte
conducted tests on pressurised cylindrical vessels for the
hydraulics associated with the large fountains in Versailles
built for Louis XIV. Mariotte observed the proportionality

103

between the pressure applied and the circumferential


extension and identified that failure occurred at certain
extensions.
Linear elastic thin shell theory emerged as a special case
of the Theory of Elasticity and in his standard work, Love
[26] laid down the basic equations which have provided
a starting point for many investigators interested in thin
shells. Thin shells proved to be a popular area for academic
study and a rich source of doctoral projects for students both
in mathematics and engineering. The equations in general,
were complex and were not easily solved except for special
cases and often only by invoking various approximations.
Many of these approximations, developed in the early 20th
century, are quite famous.
As mentioned earlier, the advent of nuclear power in the
1950s and 1960s provided both the stimulus and the
resources to accelerate the interest in solving thin shelltype problems. The large vessels in the nuclear stations
could often be considered thin (with high radius to thickness
ratios), and they were often axisymmetric in shape (and
loading too, being dominantly pressure and thermal
loading). Discontinuity-type analysis was applied to many
components; companies developed their own in-house thin
shell programs and numerical solutions were also applied.
All of this fitted well with the DbA approach pioneered in
the nuclear industry. Major advances were commonplace.
For example, the paper on nozzles in spheres by Leckie and
Penny [27] in 1963 was a landmark paper. Typical results
are given for the simplest case in Fig. 30 and these with
others were incorporated into BS 5500.
Prior to that time when the design office asked for
information on the stresses in a particular nozzle, it involved
several days work with some approximate calculations,
comparisons with our own experimental database, interpolation or extrapolation plus some guesswork. Suddenly, we
could give an answer on the telephone in a few minutes.
There were weaknesses of course in the Leckie and Penny
approach; because of the volume of information,
they ignored the stresses in the nozzle even though they
were higher than those in the sphere, arguing that they were
probably artificially high. They presented only results for
the sphere. The detailed information for the nozzle stresses
(Fig. 31) was presented later by others [28] but elucidation
of this common geometric discontinuity still continues
today and the ICPVT-10 conference chairman, Professor
Zeman has commented on the situation [29].
Of course, there were many other types of analysis being
employed. These included both experimental analysis and
other theoretical approaches. Plasticity theory received a
great deal of attention and other areas of mechanics were not
neglected. However, the codes and standards largely
remained dependent on linear elastic calculational methods,
albeit some of the Design-by-Formula routines were based
on limit loads. In BS 5500, the work of Leckie and Payne
[30] on shakedown and limit loads on nozzles in spheres
was incorporated, although in practice rarely used.

104

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 30. LeckiePenny nozzle/sphere results. R nozzle radius a sphere radius tc nozzle thickness ts sphere thickness

Fracture mechanics developed, initially at least, almost


as a separate discipline. The spectacular Liberty Ship
failures were initially puzzling (Fig. 32). These new allwelded constructions spawned a large number of failures. It
is reported [31] that 145 of these ships broke in two and over
200 suffered serious failures. Failures occurred under low
load conditions, some even in harbour situations. It became
evident that the combination of low temperature and high
restraint stresses such as those that arise at flaws or cracks
was crucial. Understanding the stress field behaviour local
to a crack tip was vital to the development of fracture
mechanics methods.

The idea of characterising the state of stress at a point by


an equivalent stress (Von Mises or Tresca usually) in order
to assess the onset of yielding was well understood.
However, it was a major step to characterise the stress
field around a crack tip in order to assess the possible onset
of fracture. Irwins crack tip stress intensity K; Wells
crack tip opening displacement, or the J integral from Rice
and Hutchinson and others have been powerful tools in the
workshop of fracture mechanics.
Then came the finite element (FE) method [32] enabled
by the advent of the digital computer. Initially known as
matrix methods and employed in the aircraft industry [33], it

Fig. 31. Maximum stresses in the nozzle.

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 32. Liberty Ship Schenectady.

developed rapidly on both sides of the Atlantic. A large


number of small-scale FE programmes based in university
departments have now given way to fewer large-scale
commercial packages. These can now handle almost any
type of constitutive law and almost any type of problem, at
least in principle. Elastic, thermal, elastic plastic, limit
conditions, buckling, creep deformation and damage,
cracks, cyclic loading, creep fatigue, plus many other
types of applications can all be accommodated. However,
the more complex the application, the more careful one
must be, since it is rather easy to model the physical reality
incorrectly and the material behaviour may be more
complex than can be described in the model. Many serious
human errors have been made.
Fatigue still seems capable of catching designers
unawares. Design for cyclic loading and fatigue can
sometimes be overlooked or neglected where it is not

105

obvious that it is a major consideration. Pressure equipment


design requires basic scantlings to be evaluated for
particular load conditions against a standard design stress.
In the case of fatigue, however, this involves having more
basic material test data than with other types of analysis
(other than creep). For many years, the conventional
approach was based on results from small-scale smooth
specimens with stress concentration factors or fatigue
strength reduction factors applied for discontinuities and
welds. Various lessons have been learned the hard way.
One of the authors (JS) was in industry when BS 1515
was introduced. This allowed higher design stresses relative
to the yield stress and consequently thinner components. It
automatically encouraged the use of steels with a higher
yield to ultimate ratio. Several years later, new problems
were experienced with fatigue cracks in boiler drum
nozzles. The higher yield strength material has no greater
fatigue resistance and the higher stresses simply mean a
shorter fatigue life [34]. Fatigue strength does not depend on
yield. Rather it depends on the ultimate stress; however, in
the presence of stress raisers, this dependence is reduced and
in the case of crack-like flaws such as in many weld details,
the dependency is virtually eliminated [35], Fig. 33.
Care should be exercised since a similar scenario could
develop with the new EN 13445 where it is allowable to use
a factor on the UTS of 1.875 for pressure action in the Direct
Route for DbA compared with 2.4 elsewhere, and 2.35 in PD
5500 or 4 in ASME VIII Div 1. Again this encourages the
use of material with a high yield to ultimate ratio. Where the
design is governed by the UTS, this would result in a thinner
vessel for EN 13445 with consequent higher stresses. If
fatigue is relevant it will be correspondingly more important
and the vessel liable to fail in a lower number of cycles than
if it had been designed to PD5500 or ASME.

Fig. 33. Material effect on fatigue.

106

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 34. Stress raising effect on fatigue.

In the late 1980s in the UK, it was considered that the


conventional approach to fatigue in the pressure vessel
standards was inappropriate for welded pressure vessels.
Broadly, it related to the process of crack initiation whereas
the fatigue life of welded structures is dominated by crack
propagation. Various factors which affect fatigue life can
have quite different effects during each process [35].
Essentially welds tend to contain crack-like flaws large
enough to by pass the initiation process. The fallacy in the
standard approach of using stress concentration factors with
fatigue curves derived from smooth specimens can be seen
in Fig. 34. The plates with the hole and the welded detail

both have a nominal stress concentration factor of 3.0.


Applying the same reduction factor for these two cases is
clearly wrong. Furthermore, the fact that the slopes are quite
different indicates that a single factor is insufficient. These
are lessons that engineers have been slow to learn.
Because of these difficulties, the section on fatigue in the
BS 5500 was subsequently changed to one which classified
weld details and employed different fatigue curves according
to the weld category. This approach had previously been used
in the bridge code BS 5400 and the offshore rules for welded
structures. A similar type of approach has been included in
EN 13445. It has been pointed out by Maddox [36] that

Fig. 35. Possible unconservatism in EN13445 fatigue approach.

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

107

the curves in EN 13445 for unwelded material are still based


on small-scale defect-free specimens and include an
allowance for material strength. He has reported that this
seems to be unconservative when compared with actual
pressure vessel failures, Fig. 35. However, caution is
required since it is not entirely clear that all of the necessary
information (e.g. geometric deviations from manufacture
and welding) is available to conduct the EN13445 analysis to
the necessary accuracy. Rauscher has issued a similar caution
when considering how to deal with surface finish on surfaces
without mechanical preparation [37]. No doubt these matters
will need further attention and clarification in the future.
6.1. Design
Analysis is usually conducted with a view to design or
assessment. In the 1960s, many impressive stress analysis
reports were produced on complex components, usually
experimental reports since there was little relevant theory.
Management had a bad habit of not being impressed by the
graphs and tables of results but simply asked the question Is
it OK? or sometimes, Can we go ahead and build it? It
was much easier to report the stress levels than to answer
those questions. However these are the questions of design.
Put another way How high can we allow the load or stress
to be? There is no simplistic answer and experience is a
great teacher. Working to a standard does not absolve the
designer from thinking or taking responsibility. A good
designer develops a sixth sense, a feel. Sometimes it is
said, If it looks right, it probably is right but that is only
true to the skilled eye.
Many of the inefficiencies have been squeezed out of the
design process. So today in fact for many components
whether we work on maximum stress, shakedown, limit
load, DbA stress categories, it does not make much
difference. Gentle increases in design stress may be possible
or small reductions in design factors but each step needs to be
taken with great care (as above in the examples of fatigue).
Modern analysis is capable of many impressive achievements. There has clearly been a cycle of ideas, indeed some
quite brilliant ideas, e.g. thin shells, shakedown theorems
based on virtual self-equilibrating stress fields, bounds
on limit loads, fracture characterisation parameters,
finite elements, weld classification in fatigue, etc.
Mistakes were sometimes made in development but then
corrected and further developed. The standards themselves
were then greatly improved thus completing the cycle.
6.2. Case study: prestressed concrete containment vessel
Recently one of the authors (JS) was involved with
Magnox BNFL on a modification to the Wylfa power station
main pressure vessel. Among other things, this involved
detailed stress and failure analysis of the concrete pressure
vessel including the simulation of penetrations, reinforcing
bar, pretensioning tendons, thermal and pressure loads, etc.

Fig. 36. 1/4 Scale model of containment vessel.

with a view to determining the effect of inserting new


expansion-type bolted anchors into the vessel. It provided
an insight into the power and capability of modern analysis.
Unfortunately, that work is not published. However a
similar example of a comprehensive illustration of the
power of modern analysis has been given recently by Prinja
and Shepherd [38]. They describe numerical analysis work
performed by NNC as part of a round robin aimed at
predicting the limit load of a 1/4 scale Prestressed Concrete
Containment Vessel which was tested at Sandia National
Laboratories in USA, Fig. 36. The vessel was a scale model
of the containment of unit 3 of Ohi Nuclear Station in Japan.
The design pressure Pd was 0.39 MPa. The model included
all the main structural features, namely a steel lined
prestressed concrete cylinder 10.4 m diameter by 10.75 m
high with a top hemispherical dome of 5.7 m radius and two
vertical buttresses. Post-tensioning takes place through 108
hoop and 90 vertical steel tendons. A 3500 mm thick
concrete slab formed the base. In addition, the main
penetrations were modelled.
The aim was to predict the response to 55 output locations
on the test vessel. Seventeen participants from USA, Canada,
France, Japan, Korea, UK, Spain, China, India and Russia
submitted pre-test analysis results. There was general
agreement up to the design pressure Pd. However, at higher
pressures, where material and structural non-linearity
influenced the response there were significant differences.
The NNC model used ABAQUS but when concrete cracking
became significant, the material model in ABAQUS was
insufficient to predict the response. A limit state test was
conducted with pressurised gas which reached 3.3 Pd before
excessive leakage due to liner tearing. Four of the 17 analyses
successfully predicted this failure mode. A subsequent
hydraulic test was taken to global collapse at 3.65 Pd, Fig. 37.
Post-test analyses were undertaken. The limitations in the
ABAQUS concrete material model only allowed analysis up
to 1.54 Pd. A special concrete material model ANACAP
developed by the ANATECH Corporation was updated into
the ABAQUS FE model. This allowed the analysis to be

108

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 37. Comparison of finite element analysis and actual failure.

continued to final collapse at a pressure of 4.2 Pd. This


exceeded the actual test value but the general behaviour and
failure location was well predicted. Various cautions and
recommendations are given in Ref. [38] but the work gives a
superb demonstration of the power of modern analysis
provided the analysis is conducted by someone with the
appropriate experience. A short video clip is available online,
courtesy of NNC, at www.psg.strath.ac.uk/videos.
7. Milestone: the European Pressure Equipment
Directive
The European Parliament and the European Council
formally approved the Pressure Equipment Directive [24] in
May 1997. This Directive harmonizes the laws of the 15
member countries of the European Union relating to
pressure equipment. It is part of the programme for the
elimination of barriers to trade. As such, the purpose of the
PED is to ensure the free placing on the market and putting
into service of the equipment concerned. At the same time,
it permits a flexible regulatory environment and allows
European industry to develop new techniques, thus increasing the international competitiveness. The Directive covers
a wide range of products.
The timetable for the implementation of the PED was as
follows:

May 1997: Approved by Council of Ministers


November 1999: Introduced as Law in Member States
May 2002: Obligatory in the European Union
November 1999 to May 2002: Transition period
(During this period, manufacturers were expected to
move to the PED from their own national legislation).

The Directive defines the Essential Safety Requirements regarding safety and other provisions without
reducing existing levels of protection within the community.
Compliance with the Essential Safety Requirements is
necessary under the Directive in order to ensure the safety of
the equipment or assemblies. A number of conformity
assessment procedures are available to the manufacturer by
which conformity with the Directive can be declared and the
CE mark can then be affixed to the product.
The European harmonized standards are one means of
conforming to the essential requirements of a directive. A
harmonized standard is a technical specification adopted by
the European Committee on Standardization (CEN) and by
CENELEC. The European Commission has given CEN a
mandate to undertake a programme of work in the field of
pressure equipment. CEN is recognised formally by the
Commission as being competent to adopt harmonized
standards at the request of the Commission. Harmonized
European Standards are not mandatory and other standards
may co-exist alongside them without the presumption of
conformity. National standards are to be withdrawn on
publication of the relevant CEN standard. The programme
of European Standards in the field of pressure equipment is
enormous and covers approximately 800 adopted and draft
Standards. There are about 300 candidate harmonized
product standards, 250 candidate harmonized materials
standards and 250 harmonized supporting standards covering welding and non-destructive examination.
The principal committees and product standards are:

CEN/TC54 Unfired Pressure Vessels


CEN/TC69 Valves
CEN/TC74 Flanges
CEN/TC267 Industrial Piping
CEN/TC269 Shell and Water Tube Boilers

Work on the Unfired Pressure Vessel Standard was


started well in advance of the Approval of the Council of
Ministers in order that there would not be a long gap
between approval and issue of the standard. EN 13445 was
issued in May 2002. It is now undergoing minor corrections
through the Migration Help Desk which is being coordinated by France (http://www.unm.asso.fr/en13445 en/).
Other standards have been released namely:
EN 12952 Water-tube boilers and auxiliary installations
EN 12953 Shell boilers
EN 13480 Metallic industrial piping
and there are others on the way.

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

109

It is early days yet but almost certainly time will establish


that the PED has been a major landmark in Pressure Vessel
Technology.

8. Milestone: Risk-Based Inspection Assessment and


Maintenance
The integration of the technologies associated with
pressure systems to embrace risk assessment, appropriate
inspection, analyses and assessment followed by selective
maintenance is a real opportunity. The approach can be
applied from the design stage through manufacture, inspection, repair and replacement to life extension. An integrated
approach ought to lead to increased reliability and safety with
greater economy in cost and effort [39]. The simplest
approach to maintenance is to wait until a component fails
before taking action. Many of our early motorcars may have
been operated in this way. Often it was highly effective but
sometimes not, although the risk element was not high.
One of the authors (JS) was introduced to the idea of
planned maintenance or scheduled maintenance when he
was given the job of introducing such a scheme in a Pilger
Tube Mill in Stewarts and Lloyds of Glasgow in 1959. The
cost of down time in the Pilger Mill was high. In the process
of seeking to understand the failure cycle in critical
components, it was possible to eliminate some of the causes
so that the process of introducing the system realised
benefits even before it operated.
The next step is risk-based assessment linked to
maintenance. The idea of risk-based assessment is to
reduce the overall risk exposure on plant. This is achieved by
focussing inspection, assessment and maintenance on those
locations, where the risk is perceived to be high. Risk can be
defined as the product of likelihood of an unexpected failure
and the consequence of the event. The approach is sometimes
termed risk-based inspection. Fig. 38 shows the concept.
ASME has supported the development of more formalised approaches to the subject whereas in the United
Kingdom, custom and practice have dictated the pace. The
methodology results in a set of recommendations on how
much preventative effort is required. The inspection and the
subsequent analysis of the data quantify the condition of the
item by determining the likelihood of failure and calls for
corrective action if necessary. In this way it reduces the
likelihood of an unexpected failure. A good summary is
given by Jovanovic [40].
An important issue is that of failure consequence.
Various attempts have been made to identify risk categories.
The system used by Shell is a good example [41].
There are three categories:
Category A functional failure of this type represents a
potential health, safety and/or health risk as well as
disrupting production in many cases.

Fig. 38. Risk-based inspection approach.

Category B where functional failure is likely to impact


availability and therefore production or operating costs.
Category C where functional failure affects neither
health and safety nor short-term operations. Limited
loss of functionality and the only direct cost is repair.
The interpretation and associated examples will vary
according to the industry concerned. As the plant will
generally contain pressure vessels, the implications of RBA
will involve current pressure vessel technology. Article 10
of the PED requires the classification of pressure equipment.
Manufacturers need to classify them as either low risk
equipment (sound engineering practice) or into one of the
four conformity assessment categories I IV according to
the tables in Annex II of the PED. The categories are
dictated by the type of equipment, the state of the fluid
contents, a classified fluid group and the pressure/volume of
the equipment. Based on this information, the appropriate
classification can be found and relevant conformity assessment modules identified.
The risk-based approach simply provides a regulated
approach to ensuring plant is safe. The elements of the
methodology are technical approaches well established in
the relevant fields, which include inspection, analysis,
monitoring and materials behaviour. Risk is two-dimensional, combining probability of occurrence with the likely
consequence. Analysis of risk may be quantitative or
qualitative. An example of the latter is given in Ref. [42]
with bands of probability and consequence. The decision to
replace or repair a component may need to be justified
financially if the consequence is not extreme. Thus,

110

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

the plant was in a fit state to enjoy a reasonable life


extension beyond its original 25-year operating life. At that
time Scotland had excess capacity and it was not an
economical proposition to keep it operating. More recently
Hinkley Point A nuclear station was diagnosed as having
sufficient radiation damage to make operation on the upper
shelf toughness difficult. A solution was devised of selective
insulation to alter the operational temperature distribution to
restore upper shelf conditions. Trials showed that it was
possible to install the insulation remotely through several of
the top cap nozzles. However, an economic decision was
taken to shut the plant down.
Fig. 39. Early X-ray machine.

the whole idea of asset management as an over arching


concept enters the field. However, for any of this to be
successful in complex plant, good analyses methods and
appropriate inspection procedures are essential. The former
has already been touched on and the latter will now be
briefly discussed.
Non-destructive examination has come a long way in a
short time. Fig. 39 shows an early X-ray machine in
operation at Babcock & Wilcox, Renfrew in the late 1950s.
It is a 2 million-V machine supported on an overhead
travelling pantograph operating in a 20 m 23 m chamber.
The chamber was approximately 1 m thick concrete walls
and contained a 100 ton turning gear system for drums. In
the intervening years, great strides have been made and
NDE is now highly advanced both organisationally and
technically, but it must be carefully managed if the
necessary quality is to be achieved [43]. The traditional
role of NDT during manufacture has been extended to
embrace the increasingly important in-service inspections
and this is especially so in ageing or extended life plant.
Many methods are available including visual surface and
volumetric methods. Ultrasonics, eddy current or magnetic
flux leakage are used on plates, forgings and pipes. Welds
are inspected by magnetic particle for ferritic steels or dye
penetrant for austenitics as well as by the volumetric
systems of radiography or ultrasonics. The trend is to
automated ultrasonics so that a permanent record can be
made (an earlier weakness).
However, the history of NDT is not without problems
[44]. The PISC series of round robin trials revealed some of
the difficulties that can be encountered viz. failure to find or
size quite large known defects or vice versa. Roy Nichols
was a prime mover in some of this work. In the PISC II
programme, he was the chairman of the International Board,
and remained chairman for PISC III. The quality of the
process and the infrastructure for NDT is paramount. When
well controlled, its achievements are impressive.
One of the authors (JS) was involved with two
assessments on major power plants where economics
overruled the technical assessment. The Hunterston A
nuclear station long term safety review concluded that

8.1. Case study: repair of heat exchanger on Sizewell A


station
The boilers at Sizewell Power Stations referred to earlier
entered service in 1966 and performed satisfactorily. Thirty
years later in 1996 at the Periodic Safety Review, during
magnetic particle inspection, evidence of defects was
picked up [45]. Subsequent ultrasonic inspection revealed
a range of defects in three out of the eight boilers. In all,
about 2 miles of weld was inspected. The worst situation
was in boiler 2C, where there was a sub surface defect up to
25 mm in depth extending over 4 m in the circumferential
direction in the HAZ around the weld between two courses
(6 and 7). This was the same boiler that had suffered the
catastrophic failure during proof hydraulic test in 1966.
Extensive diagnostic work concluded that the cracking had
been present since the boilers entered service and that the
defects were due to the stress relief heat treatment of the
original construction. It was further concluded that, subject
to certain safeguards, a repair was feasible. It was decided to
conduct a Post Weld Heat Treatment repair. The defective
area was excavated as a groove of varying depth with a
special machine developed for the purpose and a team of
specially trained welders restored the welded area manually,
Fig. 40. An ultrasonic NDT fingerprint was taken before and
after the repair. Hot NDT techniques had to be developed so
that monitoring could take place during the repair. The final
heat treatment took place over an axial length of ca. 5 m
compared with the original site heat treatment length of
0.5 m. The repair was completed satisfactorily and the
boilers returned to service in 2000. This provides a good
example of the capability of the Nuclear Industry to
undertake such repairs to an appropriately high standard.
A short video clip of the repair procedure is available online,
courtesy of BNFL/Mitsui Babcock, at www.psg.strath.ac.
uk/videos.
In 1997, a similar but smaller defect was discovered in a
heat exchanger at Chapelcross Nuclear Power Station in
Scotland. Again, it proved to be a pre-existing crack from
the original construction 40 years before. In this case, a
complete slot, circa 250 mm long, was cut in the vessel and
successfully filled with weld material without any subsequent Post-Weld Heat Treatment [46].

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 40. Sizewell excavation and repair.

9. The next milestone: fast reactors and hydrogen


It is interesting to speculate what might provide the next
major influence or stimulus in the area of pressure vessel
technology. Of course there have been, and there will be,
many other milestones in areas that affect society without
necessarily impinging on pressure vessel technology.
However, in the pressure vessel technology domain there
are two fairly obvious candidates, namely Fast Reactors
and the so-called Hydrogen Economy either one of which
might qualify. Neither of these are new ideas but to date
they have not been seriously exploited. It seems likely that
both could be important in the future but a scenario, which
brings the two candidates together, could prove to be a
milestone-type development.
9.1. Existing fast reactor systems
Fast reactor technology is relatively well understood
[47]. As mentioned earlier, demonstration reactors and
prototype reactors and, indeed, a few full-scale civil reactors
have been built but development has been slow for a variety
of reasons. In fact the first nuclear generated electricity
came from a fast reactor in 1951 at Argonne National
Laboratory in Idaho, the EBR-1. Only 35 years later, the
worlds largest fast reactor Superphenix at 1200 MW was
built in France and reached full power in 1986. Less than 20
machines have been built in the last 50 years (the main ones

111

are listed in Table 1). Most of these have been experimental,


almost all have experienced difficulties and virtually none
has been commercially successful. Nevertheless the attractiveness of the fast breeder reactor (FBR) is clear. Natural
uranium contains less than 1% (0.7%) of the isotope U-235,
which is the fuel used in current reactor systems. In a FBR, it
is possible to convert the remaining 99% (U-238) into a
useable fissile fuel, plutonium. In this way, the existing
natural uranium could be made to produce energy at least 50
times greater than could be achieved in current reactor
systems. Hence, the reason some scientists believe that
FBRs (together with fusion reactors) are the only large-scale
energy sources viable in the long term. In March this year,
India announced that it planned to generate by mainly
nuclear means 20,000 MW by 2020 of which 2500 MW will
be through five fast reactors. The first eight will be water
reactors (by 2008) and the next will be a fast reactor. India
considers itself self-sufficient in fast reactor technology.
In current reactor systems, the neutrons causing fission
have to be slowed down by use of a moderator to the
appropriate energy level to allow fission to take place. In
contrast, in the fast reactor, the velocity of the neutrons
must remain fast (close to their initial release rate) in order to
maximise the production of plutonium. Materials which act
as moderators are, therefore, undesirable; unfortunately this
includes water. The favoured coolant (being a poor
moderator) is liquid metal and liquid sodium has been the
main choice (melting point 100 8C, boiling point 900 8C).
There are other possibilities but the Liquid Metal Fast
Breeder Reactor (LMFBR) is the most promising so far.
Sodium has excellent heat transfer properties but unfortunately it reacts explosively if it comes into contact with air or
water. It also tends to pick up radioactivity. Thus, all
LMFBRs have a secondary sodium loop (non-radioactive)
which carries the heat from the primary circuit via an
intermediate heat exchanger (IHX) to the steam generator.
There are two basic systems known as the pool and the
loop configurations each with its own advantages and
disadvantages. Fig. 41 shows a typical pool-type reactor that
has the core, the primary circuit, the pump and the IHX inside
a large (site fabricated) vessel. The loop-type has a smaller
vessel housing the core with an external IHX and pump. The
latter makes for easier fabrication, inspection and maintenance although additional shielding is required for the primary
loops. A typical loop-type layout is shown in Fig. 42. The
integrity of the sodium pipework in either case is paramount.
One of the authors (JS) had research interests for many
years in high-temperature pipework, where creep is present
and one of the main applications was to LMFBR systems.
The typically tight configurations of such systems coupled
with the temperature regimes require a full understanding of
the flexibility and creep behaviour of the piping. He was
privileged during the 1970s and until the mid 1980s to work
with colleagues in Japan (PNC), France (CEA) and USA
(Westinghouse and Brookhaven National Laboratories) as
well as with the UKAEA and NNC [48 51] and for a period

112

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Centre of PNC have conducted a large research programme


on bellows units [54]. Bellows result in a tighter layout,
Fig. 45, but ensuring the integrity of the bellows
components is even more difficult than ensuring the
integrity of the pipe bends.
Of course, there are many difficulties associated with fast
reactors. There are technical ones in terms of long-term safe
operation, and public perception ones, since plutonium is
the stuff of bombs as well as crucially, commercial problems
in driving down costs to make the technology robustly
competitive. However, as energy requirements rise and
existing sources become depleted, costs may stabilise at
feasible rates.
9.2. Hydrogen

Fig. 41. Pool-type fast reactor.

acted as consultant to the US Department of Energy (DoE).


He was also invited to conduct a survey of work on fast
reactor piping on behalf of the European Community [52].
This work and that of others concentrated on the pipe-bend
as a means of ensuring the necessary flexibility. Fig. 43
shows a test on a simple pipe-bend at PNC [53].
In Japan they have also seriously investigated the use of
flexible bellows, Fig. 44. Much good work has been done in
Japan on bellows going back to the original research work of
Professor M. Hamada. More recently, O-arai Engineering

The authors have little knowledge of hydrogen as a fuel. In


the late 1960s there was some interest in the hydrogen
economy in the UK. A large-scale project was formulated
and discussed with the then DoE. At that time the advice
given was that there would be no backing from DoE since
the UK had all the energy it required in the North Sea. There
ended any involvement with hydrogen research! Others have
not been so short sighted. Demonstrations of the possibilities
of using hydrogen as a fuel, especially as a portable energy
source for commercial vehicles, are available in several
countries. A number of manufacturers are developing
hydrogen technology for cars. General Motors is undertaking
a project (HydroGen3), where standard Zafira people carriers
with hydrogen fuel cells will operate in Tokyo for one year
(as delivery vehicles for FedEx). Performance and range are
modest but practical. HydroGen3 stores hydrogen in tanks
pressurised to 700 bar (circa 70 MPa). The attraction of zero
emissions is clear. An interesting major project is underway
in Iceland, established by an Industry/Government consortium that includes Shell, Daimler-Chrysler, Norsk Hydro and
several Icelandic companies [55]. Icelands first hydrogen
fuel station is located just outside Reykjavik. A fleet of buses
with fuel cells (Fig. 46) will operate from this base. The
difference between the one in Reykjavik and earlier ones in
Europe, Japan or USA, is the way the hydrogen is made.
Typically hydrogen has been re-formed from conventional

Fig. 42. Loop-type fast reactor.

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 43. Pipe-bend test at PNC.

113

fossil fuels. Iceland is using electrolysis by applying a highvoltage charge to water. In the long-term, many believe that
the combination of nuclear power with electrolysis of
hydrogen will provide a reasonable and clean business
solution to providing fuel for future transport. Of course,
hydrogen is a severe fire hazard and that will need to be borne
in mind and addressed.
Storage of hydrogen also poses certain problems and
various solutions are possible. One version already mentioned is to store the fuel in pressurised tanks. Pressures of
5000 psig (34.5 MPa) are currently being suggested and
employed with 10,000 psig (69 MPa) being the likely
standard in the future. There is much to be done. The US
Department of Energy ($1.2 billion) and Daimler-Chrysler
($1 billion) have already allocated major research budgets.
9.3. Nuclear fusion

Fig. 44. Vertical sodium bellows (PNC).

Fusion needs to be mentioned. As a major energy


sourcem, it is further ahead in time and is likely to be the
next but one milestone. Nevertheless Europe, Canada, Japan
and Russia have been working in an unprecedented
international co-operation aimed at the next major step in
nuclear fusion. The ITER project (the way in Latin) seeks
to demonstrate stable production of electrical power from a
Tokamak-type design in a single major development step,
Fig. 47. The JET (Joint European Torus) in England has
generated a 2 s 15 MW pulse; the aim is to demonstrate that
up to 500 MW can be generated, hopefully in steady state,
so the single step experiment is truly gigantic. China has
joined the project in 2003 and USA has rejoined after
withdrawal in 1998. The design was finalised in 2001 and
procurement and research plans produced in late 2002. At
the meeting in St Petersburg in February 2003 when China
and USA came on board, four potential sites were identified
that met the necessary criteria. These are Cadarache
(France), Clarington (Canada), Rokkasho-mura (Japan)
and Vandellos (Spain). This is typically a $10 billion
project, half of which will be the capital installation and half
will be the estimated running costs for an initial 10-year
operation.

Fig. 45. Typical piping layouts with bends and bellows.

114

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

Fig. 46. Typical fuel cell for bus.

Experience with the ITER machine will pave the way for
the next machine, which may be expected to produce
useable electrical power at a meaningful level. This is
obviously an extended and indeterminable time scale and
the topic will not be discussed further here.
9.4. Generation IV systems and the international
forum (GIF)
Fig. 48. Sizewell B power station.

Generation IV nuclear energy systems are defined as


systems that offer advantages in the areas of economics,
safety and reliability, sustainability, and proliferation
resistance and physical protection over currently deployed
commercial reactors. They are assumed to be deployable
commercially by around 2030.
The first generation is referred to as those early reactors
developed in the 1950s and 1960s. The second generation
began in the 1970s in the large commercial power plants that
are still operating today. Generation III was developed more

recently in the 1990s with a number of revolutionary designs


that offer advances in safety and economics. The UK has very
few, if any, generation III reactors. Perhaps the PWR at
Sizewell B might qualify, Figs. 48 and 49. This was the UKs
first PWR. It was expensive both in time and cost. It was
preceded by a protracted public inquiry and very strict safety
requirements were imposed. Beyond 2030, the prospect for
innovative advances through renewed R & D has stimulated

Fig. 47. ITER project reactor schematic.

Fig. 49. Sizewell B reactor vessel.

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

interest worldwide in a fourth generation of nuclear energy


systems. It is considered that any expansion of current
technologies of energy production might have adverse
effects in terms of the environment and climate change. In
the future, the environmental benefits of nuclear energy
could extend to other energy products besides electricity. For
example, nuclear energy can be used to generate hydrogen
for use in petroleum refinement and as a transportation fuel to
reduce the dependence upon oil, and to desalinate water in
areas, where fresh water is in short supply.
Ten countriesArgentine, Brazil, Canada, France,
Japan, the Republic of South Africa, South Korea, Switzerland, the United Kingdom, and the United Stateshave
agreed that it is time to develop a future generation of
nuclear energy systems, known as Generation IV. The
project is facilitated and supported by Idaho National
Engineering and Environmental Laboratory (INEEL) and
Argonne National Laboratory. Ten countries have joined
together to form the Generation IV International Forum
(GIF) to develop future-generation nuclear energy systems
that can be licensed, constructed, and operated in a manner
that will provide competitively priced and reliable energy
products while satisfactorily addressing nuclear safety,
waste, proliferation, and public perception concerns. The
objective is to have such systems available for international
deployment by 2030, when many of the worlds currently
operating nuclear power plants will be at or near the end of
their operating licenses. In July 2002 in Rio de Janeiro, GIF
identified six reactor systems most likely to meet the goals
of Generation IV. These are given in Table 3.
A technology roadmap was published in December 2002
[56]. This will be a planning document and the basis for
Table 3
Generation IV systems
Generation IV System

Acronym

Deployment

Gas-cooled Fast Reactor System:


a fast neutron spectrum, helium
cooled reactor with a closed
fuel cycle
Lead-Cooled Fast Reactor System:
a lead/bismuth eutectic liquid
metal-cooled reactor with a closed
fuel cycle
Molten Salt Reactor System:
a circulatory molten salt fuel
mixture with an epithermal-spectrum
reactor
Sodium-Cooled Fast Reactor System:
a fast spectrum sodium-cooled reactor
with a closed fuel cycle
Supercritical Water-Cooled Reactor:
a high temperature, high pressure
water-cooled reactor operating above
the thermodynamic critical point of water
Very-High-Temperature Reactor System:
a graphite moderated, helium-cooled
reactor with a once-through uranium
fuel cycle

GFR

2025

LFR

2025

MSR

2025

SFR

2015

SCWR

2025

VHTR

2020

115

international research programmes. It covers the various


systems in detail and identifies the strengths and weaknesses
against an evaluation methodology. Some of these systems
could be employed earlier than the target date of 2030 and
possible implementation dates are also shown in Table 3. It
will be noted that the nearest deployment date relates to the
sodium-cooled FBR since this is the system that is best
known and with which there is the greatest experience as
described above.
Generation IV is an exciting and challenging project but it
is really just the beginning. It is estimated that each system
may require 1 billion of research. Individual countries will
only participate in the systems they favour and the viability of
each system will be constantly monitored as programmes
develop. There are challenges in materials and manufacturing
which have implications for pressure vessel technology. The
reactor outlet temperatures are all high and projected to be
between 550 and 1000 8C depending on the system.
Austenitics will serve in some cases but nickel-based
superalloys will be necessary at the higher temperatures.
Of course, there are severe challenges on the fuel handling and
reprocessing cycles and these are discussed in the roadmap.
The goals for the Generation IV Systems are given in
Table 4 and these form a significant basis for the evaluation
of the systems.
Table 4
Generation IV goals
Goals for generation IV
nuclear energy systems
Sustainability-1

Generation IV nuclear energy systems will


provide sustainable energy generation that
meets clean air objectives and promotes longterm availability of systems and effective fuel
utilization for worldwide energy production
Sustainability-2
Generation IV nuclear energy systems will
minimise and manage their nuclear waste and
notably reduce the long-term stewardship
burden, thereby improving protection for the
public health and the environment
Economics-1
Generation IV nuclear energy systems will
have a clear life-cycle cost advantage over
other energy sources
Economics-2
Generation IV nuclear energy systems will
have a level of financial risk comparable to
other energy projects
Safety and Reliability-1
Generation IV nuclear energy systems
operations will excel in safety and reliability
Safety and Reliability-2
Generation IV nuclear energy systems will
have a very low likelihood and degree of
reactor core damage
Safety and Reliability-3
Generation IV nuclear energy systems will
eliminate the need for offsite emergency
response
Proliferation Resistance
Generation IV nuclear energy systems will
and Physical Protection-1 increase the assurance that they are a very
unattractive and the least desirable route for
diversion or theft of weapons-usable materials,
and provide increased physical protection
against acts of terrorism

116

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

The concept of a Nuclear Energy park, with Generation


IV reactors, thermal reactors, reprocessing facilities, waste
disposal repositories and even combined hydrogen generation facilities is suggested. Should these come to fruition
we would truly find another milestone with major impact on
pressure vessel technology. The challenges are formidable,
some of the dangers can be anticipated but the rewards
could be great. Predictions are easy but we must hope that
the reality will be as good as the anticipation.

10. Concluding comments


The present paper has sought to identify major milestones in pressure vessel technology in the widest sense.
There have been many significant advances that have not
been covered because of lack of space and time. The
advances in materials (both behaviour and production
processes), the developments in manufacturing and fabrication as well as the improvements in joining technology have
not been considered in detail. Apologies are offered to
anyone who feels that their area of interest and expertise has
been under represented. Nevertheless, it is clear that the
achievements and capability of the pressure vessel technology community are impressive. Indeed it underpins the
basic fabric of society by ensuring secure supplies of energy
and many other commodities too numerous to mention. Too
often, this is simply taken for granted by both the public at
large and the powers that be, neither realising the level of
effort and expertise required.
The community is reasonably strong; it has great
experience and good communication channels through
technical reports and journals as well as conferences such
as ICPVT. Conferences are a prime mechanism for
networking; without them we would probably never have
met or been able to visit colleagues all over the world. It is a
pity that pressures (not psi or Pascal) in industry and
universities seem to mitigate against conference attendance.
Please keep up the good work both with ICPVT and IJPVP.
As in every generation, one worries about ageing personnel
and loss of experience. Globalisation has moved manufacture from the developed countries so that design and
research is sometimes divorced from manufacture and
practical experience not to mention the difficulty of repair or
maintenance. This could be a danger in the longer term.
Certainly key individuals and expertise are being lost to the
community in the UK and it may well be similar elsewhere.
Returning to the cycle of idea ! failure ! analysis
! maturity and/or codes and standards, it seems that the
cycle needs to be constantly restarted. Herein lies the
challenge for both the present and future engineering
communitythat of generating new ideas. From these
new ideas, the cycle will engage and the new milestones of
the future, both large and small, will emerge.
Energy has figured as a theme throughout this paper. It
will continue to be a major driver in the future. Earlier this

year, the UK Government issued a White Paper on energy. It


starts with the commitment of the developed economies to
cut greenhouse gases by 60% by 2050 and suggests that this
could be done for a cost of only 1% of GDP (in 2050).
Energy diversity is postulated with a renewables target of
20% by 2020. Currently 30% of electricity is supplied by
nuclear but by 2025 in the UK, only one nuclear plant is
expected to be operational. By 2006, the UK will be a net
importer of gas and by 2010, an importer of oil. Coal is
likely to be exhausted within 10 years and by 2020, the UK
could be dependent on imported energy for 75% of its
needs. Despite all this, the Government has made no
comments regarding future nuclear build, but it has not been
ruled out. In the short-term decommissioning, although not
a glamorous area, will offer many opportunities and
valuable experience but the long term strategy in the UK
is rather uncertain.
In the meantime, there are many challenges and
opportunities. The single market in Europe with the PED
provides opportunity to the swift. The interest from America
and the rest of the world has stimulated the International
Standards Organisation (ISO) so that trading is likely to be
facilitated not just in Europe but also across the world.
The European Pressure Equipment Research Council
(EPERC), modelled on PVRC (USA) and JPVRC (Japan),
has been a stimulus to research and technology transfer in
Europe. Tribute should be paid to Mr Guy Baylac without
whose energy and effort EPERC might never have
happened. EPERC has a number of strategic projects in
Design, Materials, Testing and Inspection, and Operation
and Maintenance (http://eperc.jrc.nl). These linked to the
sister councils in USA and Japan will help to provide a way
forward for the industry. The World Wide Web can
facilitate matters here and the concept of the Virtual
Institute currently being explored seems to be a powerful
vehicle.
The greatest threat, as in any enterprise, is complacency
and lack of vision. In the UK, now there are few university
courses in pressure vessel technology and even fewer in
nuclear energy. It may be similar elsewhere. Continuing
professional development courses in the area are increasingly poorly attended. One understands some of the
difficulties and pressures faced by commercial organisations
and academia but when todays problems prevent consideration of the future, there is likely to be a poorer future.
Expecting young engineers to tackle the complex features of
design in the pressure vessel technology area without
adequate background could be a recipe for disaster.
In summary, at present the pressure vessel technology
community is alive and well. It has been enormously
successful in responding to major milestones in world
developments. It routinely handles potentially explosive
environments and hazardous substances safely in the service
of the wider community. Roy Nichols would approve, albeit
his memory might spur us on to greater efforts and certainly
closer co-operation and collaboration.

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118

11. Tailpiece
It is well known that Roy Nichols was fond of limericks
and used them on a variety of occasions. It seems
appropriate to finish in that style.
There was a fine fellow named Roy
Who, a limerick or two did employ.
In ICPVT,
His memorial will be,
He would wish us, this conference, to enjoy.
Acknowledgements
Grateful thanks are expressed to all colleagues, friends
and organisations mentioned in the text who have
contributed material and information for this lecture.
Special thanks to John Darlaston, Nawal Prinja, Ceri
Green, Aileen Petrie and Pat Owen.
References
[1] Fletcher W. Book of famous Scots who changed the world. Glasgow:
Lang Syne Publisher; 1955. ISBN 1-85217014X.
[2] Pullin J. Progress through mechanical engineering. London: Quiller
Press Ltd; 1997. ISBN 1-899163-28-X.
[3] Timoshenko SP. History of strength of materials. New York: Dover;
1983. ISBN 0-486-61187-6.
[4] Spence J, Tooth AS. Pressure vessel design: concepts and
principles. London: E & FN Spon/Chapman & Hall; 1994. ISBN
0-419-19090-5.
[5] Nichols RW. Pressure vessel codes and standards: developments in
pressure vessel technology-5. Amsterdam: Elsevier; 1987. ISBN 185166-048-8.
[6] Weaver G, McNulty JB. An evolving concern: technology safety
and the Hartford Steam Boiler Inspection and Insurance Company.
The Hartford Steam Boiler Inspection and Insurance Company;
1991.
[7] Bernstein MD. Design criteria for boilers and pressure vessels in the
USA. Int Des Criteria Boilers Pressure Vessels, ASME PVP 1988;
152:11137.
[8] Cross CA. Power systems analysis, 2nd ed. New York: Wiley; 1986.
ISBN 0-471-86206-1.
[9] Kuijian G. 1999 Boilers, pressure vessels and piping accidents in
China. Chin Boiler Pressure Vessel Safety, Engl Ed 2000;7:348.
Also 2001 Accident report on boilers, pressure vessels, pressure
piping and special equipment. Chin Boiler Pressure Vessel Safety,
Engl Ed 2002;9:3942.
[10] Babcock M. Publicity material.
[11] BS 5500. Specification for unfired fusion welded pressure vessels
(Now PD 5500). British Standards Institution, 1st Edition 1976.
[12] EN 13445. Unfired pressure vessels. CEN 2002.
[13] BS 1515. Fusion welded pressure vessels for use in the chemical,
petroleum and allied industries 1965 (withdrawn). British Standards
Institution 1965.
[14] Spence J, Adam GR, MacKenzie AC. Hinkley Point nuclear power
stationpressure test on No. 1 reactor sphere. Report No. 1/61/24
Babcock and Wilcox; 1961.
[15] Moffat DG, Spence J. Trawsfynydd nuclear power stationpressure,
leak and vacuum tests on reactor pressure vessels. Report No. 1/62/98
Babcock and Wilcox; 1962.

117

[16] Harris H, Hafthe JS, Vaughon HG, Lees M, Burdekin FM, Wells AA.
Special report on Failure of a boiler during hydrostatic testing,
Sizewell Nuclear Power Station. Iron and Steel Institute; 1964.
[17] Burdekin FM. The role of fracture mechanics in the safety analyses of
pressure vessels. Int J Mech Sci 1982;24(4):197 208.
[18] Smith N, Hamilton IG, Failures of heavy pressure vessels during
manufacture and hydraulic testing, vol. 76. West of Scotland Iron and
Steel Institute; 1968 69. p. 111 47.
[19] On the brittle fracture of a high-pressure boiler drum at Cockenzie
Power Station. Report of the Board of Enquiry. South of Scotland
Electricity Board; January 1967.
[20] Langer BF, Merston J, Cooper WE. Tentative structural design basis
for reactor pressure vessels and directly associated components. US
Dept of the Navy Report, Washington DC; 1958
[21] ASME, ASME boiler and pressure vessel code. New York: American
Society of Mechanical Engineers, 2001.
[22] Spence J. Engineering the future: past and present perspectives.
London: Presidential Address, Institution of Mechanical Engineers;
1998.
[23] Steel WJM, Spence J, On propagating buckles and their arrest in subsea pipelines, vol. 197A. Institution of Mechanical Engineers; 1983. p.
13947.
[24] Directive 97/23/EC. On the approximation of the laws of the member
states concerning pressure equipment. Directive issued by the
European Parliament and the Council; 29 May 1997.
[25] Rossmanith HP. Fracture mechanics and materials testing; forgotten
pioneers of the early 20th century. Fatigue and fracture of engineering
materials, vol. 22. Oxford: Blackwell; 1999. p. 78197.
[26] Love AEH. The mathematical theory of elasticity, 1st ed. Cambridge:
Cambridge University Press; 1892.
[27] Leckie FA, Penny RK. Stress concentration factors for the stresses
at nozzle intersections in pressure vessels. Welding Research
Council Bulletin (No. 90), New York, USA: Welding Research
Council; 1963.
[28] Rodabaugh EC, Witt FJ, Cloud RL. Stresses at nozzles in spherical
shells loaded with pressure, moment or thrust. Phase report No. 2,
Washington, USA: US Atomic Energy Commission; 1966.
[29] Schindler S, Zeman JL. Stress concentration factors of nozzlesphere connections. Int J Pressure Vessels Piping 2003;80(2):
8796.
[30] Leckie FA, Payne DJ. Some observations on the design of spherical
pressure vessels with flush cylindrical nozzles. Proc Inst Mech Engrs,
London, 1966;180(20):497512.
[31] Brock D. Elementary engineering fracture mechanics, 4th ed.
Dordrecht: Kluwer Academic Publishers; 1991. ISBN 90-247-2580-1.
[32] Zienkiewicz OC, Cheung UK. The finite element method in
continuum mechanics. London: McGraw-Hill; 1968. (4th Edition
expanded 1984).
[33] Argyris JH. Energy theorems and structural analysis. London:
Butterworths; 1960. Reprinted from articles in Aircraft Engng
19541955.
[34] Spence J, Carlson WB. A study of nozzles in pressure vessels under
pressure fatigue loading. Proc Inst Mech Engrs, Lond 1967/68;
182(31):657 84.
[35] Maddox S. Fatigue aspects of pressure vessel design. Pressure vessel
design: concepts and principles, London: E & FN Spon/Chapman &
Hall; 1994. [chapter 9], ISBN 0-419-19090-5.
[36] Maddox S. Assessment of pressure vessel design rules on the basis of
test data. In: Banks WM, Nash DH, editors. Pressure equipment
technology: theory and practice. London: Institution of Mechanical
Engineers; 2003. p. 23748. ISBN 1-86058-401-2.
[37] Rauscher F. Fatigue of non-welded pressure vessels made of
high strength steel. Int J Pressure Vessels Piping 2003;80(3):
197204.
[38] Prinja NK, Shepherd D. Numerical simulation of limit testing of a 1/4
scale pre-stressed concrete containment vessel. In: Banks WM, Nash

118

[39]

[40]

[41]
[42]

[43]

[44]

[45]

[46]

J. Spence, D.H. Nash / International Journal of Pressure Vessels and Piping 81 (2004) 89118
DH, editors. Pressure equipment technology: theory and practice.
London: Institution of Mechanical Engineers; 2003. p. 171 97. ISBN
1-86058-401-2.
Spence J, Darlaston J. Pressure vessel technology: past present and
future. The George Stephenson lecture 1998/99, London: Institution
of Mechanical Engineers; 1998.
Jovanovic A. Risk based inspection and maintenance in power and
process plants in Europe (RIMAP). Paper 10, MPA-Safe Workshop
2001. Nucl Engng Des 2003;226(2):165 82.
Hagemeijer PM, Kerkveld A. Risk based assessment of pressure
systems. Proc Inst Mech Engrs, Lond, Part E 1988;212.
Schmidt MEC, Mauney DA. Risk based methods and the ASME
handbook. Transaction of Conference, assuring its safe, London:
Institution of Mechanical Engineers; 1998. 1998-6 paper C535/029/
98, p. 32332.
Farley JM. Evolution of the NDT quality infrastructure for
pressure vessels and equipment. In: Banks WM, Nash DH, editors.
Pressure equipment technology: theory and practice. London:
Institution of Mechanical Engineers; 2003. p. 26980. ISBN 186058-401-2.
Farley JM. Quality and reliability through NDT. Transaction of
Conference, assuring its safe, Institution of Mechanical Engineers;
1998. 1998-6 paper C535/003/98, p. 129 38.
Exworthy LF, Little WJ, Flewitt PEJ. Diagnosis of cracking in the
boiler shell seam welds at Sizewell A Power Station. Int J Pressure
Vessels Piping 2002;79:413 26.
Wintle J, Jones R. Teamwork with industrynuclear boiler repaired
using multi-disciplinary approach. TWI Bull 2000.

[47] McHugh G, Merrick AR. The fast-neutron breeder fusion reactor.


London: The Royal Society; 1990. ISBN 0-85403-412-9. Pub also in
Philos Trans R Soc Lond, Ser A;331(1619):287453.
[48] Spence J. Creep analyses of smooth curved pipes under in-plane
bending. J Mech Engng Sci 1973;15(4):25265.
[49] Spence J. Aspects of pipework design in the creep regime. Institution
of Mechanical Engineers Conferenc on Creep and Piping, Pub. No. 5;
1974. paper C48/74.
[50] Boyle JT, Spence J. An analysis for out of plane bending on curved
pipes in creep: development of method and design factors. In: Corum
J, editor. International Working Group on Fast Reactors, Specialists
Meeting on High Temperature Structural Design Technology,
Champion, Pennsylvania. IAEA; April 1976. p. 14969.
[51] Spence J, Boyle JT. Simplified methods for piping analysis in the
creep regime. Institution of Mechanical Engineers Conference on
Recent Developments in Higher Temperature Design Methods,
London; 1977. paper C275/77, p. 718.
[52] Boyle JT, Spence J. A state of the art review of secondary piping
(static and dynamic) analysis methods, with particular application to
LMBFR. Reports to European Community; Nov 1983.
[53] Imazu A. Private communication, O-arai Engineering Centre, PNC.
[54] Research and development activities on piping bellows expansion
joint for FBR. O-arai Engineering Centre, Power Reactor and Nuclear
Fuel Development Corporation, (PNC), N9530 88 010T. 1989
[55] Eisenstein P. Kicking the oil habit. Prof Engng, Inst Mech Engrs,
Lond 2003;16(8):267.
[56] A technology roadmap for generation IV nuclear energy systems. US
Dept of Energy Research and Advisory Committee and the
Generation IV Int. Forum; Dec 2002. GIF-002-00.

Вам также может понравиться